column name 'treat' is not found

[Netanel-bp]

Hello,

I have been trying to use the triplediff package. My command is:

att_gt <- ddd(
  yname = "turnout", 
  tname = "year", 
  idname = "did", 
  gname = "year_no", 
  pname = "part_q", 
  xformla = ~cov1,
  data = df, 
  control_group = "notyettreated", 
  base_period = "universal", 
  est_method = "dr"
)

Whatever I do, I cannot run the function because I always get this error message:

Error in `[.data.table`(small_groups, , treat) : 
  j (the 2nd argument inside [...]) is a single symbol but column name 'treat' is not found. If you intended to select columns using a variable in calling scope, please try DT[, ..treat]. The .. prefix conveys one-level-up similar to a file system path.

After some consultation, I believe this is not a mistake from my end, by possibly a bug in the package. Is that the case?

Thank you very much,
Netanel

[pedrohcgs]

Thanks a lot for this. Would it be possible to share the data (or just a subset of it so we can replicate it?

…

---------------------------------------------------- *Pedro H. C. Sant'Anna* *https://psantanna.com <https://psantanna.com>*

---------------------------------------------------- Warning: This email may contain confidential or privileged information intended only for the use of the individual or entity to whom it is addressed. If you are not the intended recipient, please understand that any disclosure, copying, distribution, or use of the contents of this email is strictly prohibited. ----------------------------------------------------

On Mon, Oct 20, 2025 at 10:03 Netanel-bp ***@***.***> wrote: *Netanel-bp* created an issue (marcelortizv/triplediff#29) <#29> Hello, I have been trying to use the triplediff package. My command is: att_gt <- ddd( yname = "turnout", tname = "year", idname = "did", gname = "year_no", pname = "part_q", xformla = ~cov1, data = df, control_group = "notyettreated", base_period = "universal", est_method = "dr" ) Whatever I do, I cannot run the function because I always get this error message: Error in `[.data.table`(small_groups, , treat) : j (the 2nd argument inside [...]) is a single symbol but column name 'treat' is not found. If you intended to select columns using a variable in calling scope, please try DT[, ..treat]. The .. prefix conveys one-level-up similar to a file system path. After some consultation, I believe this is not a mistake from my end, by possibly a bug in the package. Is that the case? Thank you very much, Netanel — Reply to this email directly, view it on GitHub <#29>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABE734ZLVT2NUHGNX7EN2LL3YTTSPAVCNFSM6AAAAACJV3LFI2VHI2DSMVQWIX3LMV43ASLTON2WKOZTGUZTENJXHA3TKNQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[marcelortizv]

Hey @Netanel-bp, thanks for your interest in the package! I believe this issue was fixed in another issue. Could you confirm with us which version of the package you have installed on your computer?

[Netanel-bp]

Thank you both for the prompt answer and willingness to help!
Regarding the version, I updated it just before running it; the version is: ‘0.1.0’.

Regarding the data, I am attaching it here. I should note that this is just a simplified version of my data, generated merely for troubleshooting after facing some issues with the package. For example, at the moment, Cov1 is just a variable drawn randomly from a uniform distribution [0,1], but later I will want to replace it with a set of fixed effects.

Thanks a lot, I appreciate your help!
Netanel

[marcelortizv]

Thanks for sharing the data, @Netanel-bp. It seems you have the old version (with the bug) installed in your computer. Could you please update to version 0.1.2 and verify that you still having the issue?

[Netanel-bp]

Hi Marcelo.
Thanks a lot for your response!

I updated the package via running: devtools::install_github("marcelortizv/triplediff@v0.1.2") . I verified it by checking packageVersion("triplediff"). The version now is 0.1.2, but the problem seems to persist. How should I do?

Thank you very much!
Netanel

[pedrohcgs]

We are on it! Will circle back soon

[marcelortizv]

@Netanel-bp I just ran the code with the data you shared with us. It seems that the error you encountered at the beginning due to a "small groups" check does not appear anymore from my end. So, that issue was actually solved in version 0.1.2.

However, the code is currently generating warnings and early stopping due to insufficient overlap (i.e., insufficient data) in certain treated groups:

Error in value[[3L]](cond) : Error in computing regression coefficients: Subgroup 3 may have insufficient data. In addition: Warning messages: 1: In run_preprocess_multPeriods(yname = yname, tname = tname, idname = idname, : Dropped 11 units that were already treated in the first period. 2: In run_preprocess_multPeriods(yname = yname, tname = tname, idname = idname, : Be aware that there are some small groups in your dataset. Check groups: 475.

A cross-table analysis reveals that you do not have observations for groups (S=473, Q=1), (S=474, Q=0), and so on, over time. Consequently, we are unable to compute the group-time average treatment effect for that cohort (indeed this issue is also applicable to the rest of the treated cohorts. Thus, eq (4.8) in the paper cannot be computed due to the lack of this information).

We are working on a small fix to print more informative warning/error messages when the data present these problems.

[Netanel-bp]

Hi Marcelo,

This is extremely helpful! I fixed the issue, and the package ran. Thank you very much!